Nathaniel Whittemore predicts AI agents dominate workplaces

Startups (1)

• Cerebras IPO saw massive demand, debuting at a $40B valuation, hitting $100B briefly, and settling at a $66B market cap.

Markets (1)

• Jim Cramer warned investors Cerebras stock had detached from fundamentals, advising caution until a pullback.

AI & Tech (2)

• Figma revenue grew 46% last quarter, accelerating from 40%, credited to AI features; stock rose 8% after-hours.
• Codex mobile app allows full agent management from phones, shifting work from execution to continuous oversight.

Chips (1)

• Nvidia stock surged 20% over seven days, nearing a $6 trillion valuation.

Big Tech (1)

• OpenAI considers legal action against Apple for breach of contract regarding the ChatGPT integration.

Society (2)

• Wafa Bilal lived in a Chicago gallery for 30 days with a webcam-controlled paintball gun; strangers shot him 70,000 times, generating 80 million hits from 128 countries.
• Among people aged 18 to 34, 10 times as many live alone today as in 1950. Zaki speculates this solitary, transactional urban living may affect empathy, as anonymous interactions do not favor it.

Psychology (15)

• Jamil Zaki says empathy has three components: emotional empathy, cognitive empathy, and empathic concern. Different brain systems support each, and groups struggle with different flavors.
• Juliana Schroeder’s study found people dehumanize others more when reading their political opinions versus hearing them. Online text and avatars lack the cues that trigger real empathy.
• Altruism born of suffering describes how trauma can open people to caring more about others' pain. Community support after trauma is a key factor in whether individuals turn outward or inward.
• Oxytocin increases caring for insiders but decreases caring for outsiders, illustrating empathy’s parochial nature. Zaki argues we can choose to broaden our empathy beyond instinctive tribalism.
• Emil Bruno’s research shows extreme empathy for one’s own group can prevent compromise with outsiders, even if you also empathize with them. Police training centers sometimes demonstrate this.
• About 50% of oncologists report intense heartbreak delivering bad news. Empathy benefits patients but can be an occupational hazard, leading to defensive dehumanization among caregivers.
• Mark Penzer’s 1970s study showed people walked further away from a charity table if a person in a wheelchair was present, avoiding empathic triggers to sidestep guilt or donation pressure.
• Executioners studied in the American South dehumanized inmates, especially if they delivered lethal injections. Empathy can cause guilt or self-loathing, leading people to avoid it.
• Jamil Zaki’s virtual reality simulation of becoming homeless reduced dehumanization and increased support for affordable housing a month later. Immersive technology can stretch empathy.
• Acting training improves empathy more than visual arts training, according to Talia Goldstein’s studies. Narrative fiction also builds empathy, especially for groups different from the reader.
• Mark Levine’s study found Manchester United fans helped injured fellow fans but stepped over Liverpool fans. When asked to write about loving soccer instead of the team, they helped rivals.
• Leslie John argues reciprocity is instinctive in disclosure: sharing something sensitive signals trust and prompts others to reveal. Validation, not problem-solving, is the best response.
• Leslie John’s study showed a 'hurting effect': knowing others admitted to a sensitive act made people more comfortable revealing it themselves, even in a bare-bones online setting.
• Humble bragging - couching a boast with humility - is ineffective and irritating. Outright bragging is better, but requires careful timing, audience, and context to avoid sparking envy.
• Vulnerability can be weaponized, as in interrogations or scams. People in weakened states are more susceptible. Meta-disclosures about relational tension can resolve conflict collaboratively.

Philosophy (8)

• Christian List distinguishes the philosophical debate as compatibilism vs. incompatibilism, not determinism vs. indeterminism. Incompatibilists argue free will is impossible if the universe is deterministic, while compatibilists argue they can be reconciled.
• List criticizes standard compatibilist strategies for watering down free will. One camp rejects 'alternative possibilities' as necessary, while another redefines 'could have done otherwise' with counterfactuals, which he argues fails to capture the genuine openness of choice.
• List's core argument reframes the debate through levels of description. He asserts that determinism vs. indeterminism is a level-specific distinction, and high-level indeterminism (like coin flips) can be real, not just epistemic, and is compatible with low-level physical determinism.
• He argues for explanatory autonomy of higher levels, like intentional psychology, from lower physical levels. Intentional concepts (beliefs, desires) are not reducible to subintentional concepts, supported by a combinatorial argument showing high-level properties outnumber describable low-level ones.
• List proposes three criteria for free will: intentional agency, alternative possibilities at the agential level, and causal control (mental causation). Free will is a functional property requiring all three, distinct from phenomenal consciousness.
• He defends a realist theory of group agency, arguing organized collectives like firms and states can be intentional agents with beliefs and goals. This is supported by their treatment as unitary rational actors in economics and strategic studies.
• He argues free will is a necessary but not sufficient condition for moral responsibility. Additional normative and epistemic capacities are required, which is why chimpanzees might have free will but not bear moral responsibility.
• List adopts a non-physicalist stance on consciousness, arguing first-personal phenomenal facts cannot be accommodated within a purely third-personal scientific worldview, a position he sees as more radical than David Chalmers's.

AI & Tech (2)

• List suggests his functionalist criteria for free could apply to sophisticated AI systems or corporate entities if viewing them as agents becomes explanatorily indispensable, though this does not imply they are conscious.
• On AI, List says attitudinal volatility in LLMs challenges stable diachronic identity, which could disrupt traditional frameworks for responsibility attribution, forcing updates to our conceptual systems.

Physics (1)

• Sean Carroll notes a difference between coarse-graining in physics (naturally emergent) and in social groups (imposed by voting rules). List analogizes organizational structures to an operating system, a contingent 'normological constraint' that defines the supervenience relation for group attitudes.